The Hypothesis
I have developed a hypothesis about the nature of a musicality function F which defines the musical quality of an item of music.
- F is implemented within the brain of the individual musical listener.
- As a function of input audio perceived by the listener, F outputs a one-dimensional value which, as a function of time, is a measure of how “musical” the music, ie how “good” or how “strong” the music is.
- F does not determine the emotional quality of the music - emotion is multi-dimensional.
- Originally there was a protomusical language of emotion, spoken by prehistoric human ancestors, which expressed emotion similar to what music expresses, but which was not evaluated by any musicality function F.
- F then evolved in order to motivate humans to create and listen to protomusical vocalisations which are local optimums for the output value of F (“local” relative to some topogical notion of what counts as two different audio inputs being “close” to each other).
- Based on observations relating to musical pitch scales and the occurrence of nested regular beats
underlying musical rhythms, I have hypothesised that musicality is defined by the occurrence of
certain constant patterns of activity in cortical maps that process features of perceived audio.
- In particular F is defined by the amount of boundary between inactive regions and active regions of a cortical map. (For example, if a division between active and inactive was sufficient, then a tune could just have low pitch values and no high pitch values. But this would create only one boundary line between the active and inactive areas in a cortical map processing pitch, whereas if pitch values come from a scale, then every pitch value in the scale defines an active area in the cortical map which has a boundary between the inactive areas representing the gaps in the scale both above and below that pitch value.)
- A criteria based on spatial patterns and boundaries between active and inactive regions in a cortical map strongly suggests that glial cells may be involved, in particular astrocytes, as these cells have activity that is very much dependent on levels of neural activity of nearby neurons.
- We can take this glial hypothesis further and observe that a boundary between an active region and a neighbouring inactive region would be represented by the persistent contrast in the metabolic state of two neighbouring astrocytes A1 and A2 – where A1’s state related to the active region and A2’s state related to the inactive region.
The implication of the hypotheses so far is that musicality will be determined by the number of astrocyte pairs with this persistent contrast between associated active and inactive neural states.
However, in order to explain music perception, somehow, and somewhere, information about these astrocyte pairs has to be fed back into neural processing networks.
I have referred to an assumed processing of protomusical vocalisations. As modern humans, we are only aware of the emotions of music when music is actually musical. So we must further assume that the calculation of musicality suppresses the perception of protomusical emotion unless the calculated musicality value is high enough.
Like most perceptual processing that occurs in the brain, the processing of protomusical vocalisations will be purely neural. So if musicality is determined by the metabolic state of astrocytes in certain cortical networks, then that information has to be fed back into whatever cortical maps are responsible for determining the emotions expressed by protomusical vocalisations.
This raises the question of how far away the cortical maps determining musicality are from the cortical maps that determine protomusical emotion. They may be quite separate, or there may be some overlap, or perhaps they exist in different layers of the same cortical regions.
One might suppose that a simple one-dimensional signal could be propagated through purely astrocytic networks, or the general inter-cellular medium. Astrocytic networks in the brain are a thing. However, the speed at which those networks can transmit information is very slow compared to neural transmission. In the case of music, a music listener can begin to determine the musical quality of a musical item within only a few seconds, and this strongly constrains how far an astrocytic signal representing the musicality of a musical item could travel within that time constraint.
So the general implication is that if astrocytes do play an intermediate role in the determination of musicality, the final representation of that calculated quantity has to be fed back into neural networks in almost the same location.
We could suppose the existence of some specialised neuron whose job it is to send information about musicality to the cortical maps processing protomusical emotion. But if such a specialised neuron exists, and given that it would have to be unique to humans, we might expect that anatomical studies of human brains would already have discovered it.
Or, we could suppose the the neurons in the cortical maps processing protomusical emotion have dendrites that extend into the cortical maps where musicality is calculated. However dendrites are not known to have any significant length, in particular not more than 2mm in any direction in the human brain, so that idea doesn’t work very well.